COMP 4720 Project

出自Wiki @ CNLiuFeng.com

Description

The goal of this project is three-fold:

To understand the power model in CDMA based 3G networks. (i.e., how to calculate the signal noise ratios at each individual mobile user)
To understand the feedback-based power management ideas for downlink power control, as well as other multicast support in cellular networks papers
To implement a simulation model for downlink power management for multicast traffic in CDMA based 3G networks.

For 1

Downlink means from the base station to the mobile user direction. As described in class, for wireless communications, the signal-to noise ratio (SNR) affects the bit error rate a mobile user experiences; for proper data transmission, the power level used at the base station for each mobile user should be adjusted such that each user can have adequate, but not un-necessarily high, SNR for proper data reception, and minimal interference to others. Saved power can also be used to support more users.

There are power models to calculate the SNR at mobile stations in 3G cellular networks, parameters of the model in includes distance between the base station and the mobile user, as well as some others. Some are constants. You are asked to study the related papers and books, and understand these models.

The power model can be read from the following papers:

Xin Wang, Ramachandran Ramjee, Harish Viswanathan, Adaptive and predictive downlink resource management in next generation CDMA networks. Look for the 14-page version of this paper.

For 2

Feedback-based power control is used in today's cellular networks, mainly for unicast traffic. You are asked to read these power control schemes, fully understand them. The papers are:

for unicast
- Soumya Das, Sachin Ganu, Natalia Rivera, Ritabrata Roy, Performance analysis of downlink power control in CDMA systems, IEEE Sarnoff Student Symposium 2004
for multicast
- M. Chuah, Teck Hu, and W. Luo, UMTS release 99/4 airlink enhancement for supporting MBMS services
- George Xylomenos, Vasilis Tsakanikas, and George C. Polyzos, Reducing the transmission power requirements of the multimedia broadcast/multicast service
- Antonios Alexiou, Christos Bouras, Vasileios Kokkinos, Evangelos Rekkas. Power efficient radio bearer selection in MBMS multicast mode

The following two papers are more general than above on multicast support in cellular networks

Multicast scheduling in cellular data networks, IEEE Infocom 2007
Fixed/variable power multicast over heterogeneous fading channels in cellular networks, IEEE ICC 2008

For 3

The attached code is a cellular network simulator written in C/C++. The code was developed by a previous graduate student of mine on Linux. Just type "tar -xvf mobilityCC.tar", and "make". You should get an executable called "cell", and you can run with just "./cell". The execution takes about 6 to 7 minutes, which you certainly can try to shorten. I think the link-list management is inefficient right now.

This simulation model does not have 1 and 2 above built in it. It does have the model of a cell-based cellular network. And models for user mobility, as well as user bandwidth requirements and cell bandwidth management. It was developed for this paper:

Y. E. Liu and M. Han. A Cell-based Call Admission Control and Bandwidth Reservation Scheme for QoS Support in Wireless Cellular Networks. In Proceedings of the Workshop on Performance of Wireless Networks and Communication Systems (WiNCS 2005), pages 273-282, Philadelphia, PA. July 2005. The Society for Modeling and Simulation International (SCS). http://www.cs.umanitoba.ca/~yliu/MYPUB/wincs2005.pdf

You are asked to modify the code, or develop your own based on the cellular network model implemented in this code, adding 1 and 3 in it, and be able to model multicast power control schemes in such a network. The ultimate goal of this research is to come up with sound multicast power management schemes (preferably feedback-based) for optimal performance. The simulator that you develop will be used for studies toward this goal. You are encouraged to think of your own such schemes. The challenge here is to determine what kind of information should be sent back to the base station as feedbacks, and how a base station uses these feedbacks for optimal power control.

These materials are for you to get started. I will let you know in case I find more materials for this project. I am more than happy to talk with you regarding any issues or aspects of this project, at any time.

For locating the papers, you can simply use google scholar, IEEE and ACM digital libraries. If you find any more recent papers on this research subject, by all means add them to your bibliography. It will be part of the contribution of your project.

Mar 02, 2010 Announcement

For survey based projects, each student needs to read around 8 high quality papers.

The look is evaluated by the quality of the references as well.

Repeatly the abstracts is absolutely a No No.

Evaluation is based on "your" work.

classification
justification
pros and cons

Ellen's email on Mar 12, 2010

Hello,

I had a meeting with Travis regarding your COMP4720 course project.

Here are the work items at this time:

Understanding an existing radio propagation model implementation for unicast. Travis is working on this one.
Learning from the literature the difference between radio propagation model for unicast and that for multicast, and find the formulas that your project will use. This one is closely related to what Travis is doing, so you two should communicate as much as possible so to find the models we want fast.
Reading literature on 3G multicast power control. Focusing on the algorithms that have been proposed to do this so far.
Reading literature on MAC protocol design in wireless networks, so you can design a MAC protocol for uplink feedback transmissions.

By the next checkpoint (3/26), I would like to see a demo that has included an implementation of the multicast radio propagation model.

I would also like a write up that includes your findings regarding 3 and 4 above.

If those of you who are doing 3 and 4 are not sure what is expected, please schedule an appointment with me to talk about it.